Why The Google AdWords Keyword Tool Shouldn’t Be Used For SEO – A Case Study

Yep, mirrors my experiences exactly. Tool results of 15,000+, read world of 200ish. Utterly useless for SEO purposes other than extremely early stage keyword research in extreme cases of writers block.

You are not wrong here! i ranked number 1 for a term that google AdWords keyword tool said had 1,900 local searches and i got nothing close to that at all.

I have found so many incongruent breakdowns of keyword searches when people offer SEO they treat what this tool gives you as facts and base very expensive SEO strategies on such keywords.

Great article. Having just completed a laborious keyword research exercise using Google Adwords KWT it’s a crushing disappointment! Oh well. Thanks for the insight. What other tools do you recommend aside from these?

That’s such an interesting post!

So which tool is the best for keyword research In your opinion?

This was a really good read and I’ll definately go play around with the search based tool. I often wonder about Googles suggested search volumes. It often differs from other tools I use or from actual traffic for top performing keywords. Cheers

Nice explanation David. I had a similar problem to this just recently for the term “Quality Management […]” and I couldn’t quite put my finger on why it didn’t propel the client’s ranking.

Your article seems to answer that, I suspect for similar reasons. I have also had similar experiences with traffic phrases from the search based tool too – have you experienced this at all?

There seems to be an issue with broad phrase and exact match in all of the keyword research tools, and especially in that bracket of 12,000 – 20,000 searches. Possibly in the higher traffic terms too, but I never bother with those.

Good Article – I like it how you’re not just stating a theory but actual hands on experience!!

I was obsessed with Google Adwords keyword tool for all SEO campaigns until i came to know this. Thanks for your insights.

Hi David,

thx for this article, I used the keyword tool the same way and I figured out the same problems. Now it’s clear. Thx.

When I search for that keyphrase I see a local map result appearing above the number 3 organic result which means a #3 ranking is actually way down the page. How much do you think this skews the results?

A good solid follow up to this would be if the client ran PPC on that exact phrase too to see impression numbers.

Either way – this is very interesting, it’s not often you get such a black and white example. Thanks!

@Tom – fair point, but even at #4 you would be expecting a CTR of at least 1%. These figures work out as a CTR or 0.06%, that’s still around 20 times lower than what you would expect, minimum.

There is no doubt in my mind that the AdWords Keyword Tool grossly exagerates search volumes and/or is highly inaccurate.

Wow, I knew the data from the Adwords keyword tool were not very accurate, but I’ve never experienced such a big difference like this.

There must be some plausible reason for this deviation, but I don’t believe your theory that the marks are based on Google’s content network as well will hold.

Google Adwords uses a broad network of websites, including Google-owned sites, partners and AdSense-publishers. To distinguish parts of this network, Google uses several terms, including Search Network and Display Network.

Google’s definition of the Search Network does NOT include AdSense-publishers:
“Ads are targeted based on a user’s search terms. For example, if you search for “Italian coffee” on a search engine powered by Google, such as AOL.com, you’ll see related coffee ads next to the search results.”
(found here: http://adwords.google.com/support/aw/bin/answer.py?hl=en&answer=6104 )

When you are using the Adwords keyword tool for SEO-purposes it is good to know its limitations. After your example there can’t be any doubt about that.

The main two lessons should be:

1. Never trust absolute data from this tool, if you are using it for your SEO it should only be used for analyzing possible keywords and trends. Don’t place your bets on the given traffic for a single keyword.

2. If you want to optimize for traffic on a specific keyword, be aware that the Google Search Network does include many more platforms than Google Search alone. In it’s definition of the Search Network, Google gives AOL Search as an example, but you could also consider YouTube’s search function as ‘search engine powered by Google’.

Maybe somebody else has a better theory for the difference you have experienced.

@Darren – I don’t believe the Search Based Keyword Tool has exact/phrase/broad match options – it is purely exact match I think. I haven’t had many problems with it… yet, although it won’t show search volumes for a lot of keywords – I think that is Google protecting their IP.

@Huw / @Paul – Google Search Based Keyword Tool is the only thing I’d recommend for search volumes at this time. I’m based in the UK though, and Google has over 90% market share – so we tend to ignore Bing/Yahoo volumes.

HAve you seen Google Insights? Very useful, but some caveats apply here too. What’s interesting is the geographic breakdown and volumes over time.
http://www.google.com/insights/search/#

@sint – I know of a couple of websites that are considered part of the search network, but aren’t search engines. So you can’t really trust that definition. And it would explain a large amount of impressions for keywords. I have a client that is a very large website that is part of the AdSense Search network – it is NOT offfered to your average AdSense user – it is intended for smaller search engines such as AOL but it can be used by other types of websites it appears.

I’ve been looking for a post about what’s changed with the ‘new’ AdWords Keyword Tool because I’ve noticed a big difference and haven’t been as comfortable with the results I’ve been getting with this new version. Thanks, Dave for clearing some of this up.

As I am based in Ireland, it was (is) an important tool for me because I could get results for Ireland alone and compare to UK, US and all regions — I haven’t found another tool that includes Irish results.

David, the Sktool is not exact either. Plus it is outdated. I’ve tried it many times and reached this conclusion. Anyway, like every other tool, it should be taken with grain and salt. None of them are accurate But that’s just the catch, if they all were, we SEOs would have to face a lot more competition

My advice: use the Adwords tool and double check your results with the Free keywords tool from Wordtracker. Then triple check with something else, Nichebot or semrush for example (these are not for free). And then quadruple check with the sktool . And the try to check the apges ranking for the keyword and predict based on their alexa ranks the amountof traffic they get. In the end, you’ll be able to sort out if a keyword is ok or not.

For instance, on the given example, wordtracker says 20 searches a day. So, if you manage to get your site on top positions, you should get 50 to 200 searches per day for this keyword and relatives.

I think it`s relative. Did you test with other keywords or segments ?

@Mike from what I can tell Wordtracker data is based on dogpile.com and metacrawler.com so god knows why you would want to use that to estimate search volumes. Secondly Nichebot appears to use Wordtracker for its search volume data, so that’s out too. SEMrush gets its data from Google AdWords Keyword Tool – so there isn’t much point in checking that.

So in the end you are left with two tools. Google AdWords Keyword Tool, which as I just demonstrated is totally inaccurate for SEO, and the Google Search Based Keyword Tool.

Can anyone who doesn’t like the Google Search Based Keyword Tool demonstrate its inaccuracy with a case study? I honestly believe it is the most accurate tool out there for predicting Google search volumes (particularly in the UK).

@Heric – no I’ve not tested it, it just appears to correlate with getting high up in rankings only to be disappointed at the lack of traffic

Relative to what?

the sktool has just as many inaccuracies . wont display loads of high searched terms…

tbh its not in googles interest to lay open to everyone accurate numbers. Its open to too much abuse and speculative behaviors that have the potential to harm the user experience…. plus why would they want Bing being able to accurately conduct comparative research to try close the gap…

Fuckers….

This is a great article, Adwords can be good for generating keywords but not for actually seeing the amount of traffic those words receive.

@Tom – Yeah it is a shame they don’t show some keywords, which makes life a little more difficult…

Excellent article (The only thing that annoyed me was the usage of “m8″! Did you write this from your iPhone!?), and yes very disapointing that the SKtool doesnt show all keyterms. Comparing both tools with each other seems like a sensible way forward.

Not sure if people are aware but by the sounds of things the Search Based tool will be gone by the end of the month…

http://adwords.blogspot.com/2010/08/updated-keyword-tool-coming-out-of-beta.html

@Tom yeah I had read that article, I’m still hoping that by the time they close the SK Tool they will have incorporate it into the new tool *fingers crossed*

@Chris – lol sorry Chris (does lol upset you too? :S) too much chatting on IM over the last 10 years!

Excellent post. I agree there’s a problem here and I’ve never been clear about the differences between these tools. I will check sktool out more closely, but per link below, the window for using that is closing quickly, so don’t get too attached.

I posted on a related topic Saturday about how crappy the new Keyword Tool is (the current ‘beta’ version) as compared to the older interface. I did side-by-side comparison of the two versions on a given set of example keywords. The new version was virtually useless in generating relevant new keywords (and you’ve now shown that the quantities are worthless, too). Details here: http://www.webbizgeek.com/review-new-google-keyword-tool-sucks/

But even if sktool is better with volume estimates, we’ve only got about 2 more weeks to use it. Google announced last week that Keyword Tool is coming out of beta at the end of August and will permanently replace both old Keyword and existing Search Tool. Original post: http://adwords.blogspot.com/2010/08/updated-keyword-tool-coming-out-of-beta.html

“Both of these tools will be available in their current forms through the end of August, after which time their addresses will redirect to the updated Keyword Tool.”

@John – I’m hoping they are going to include the SKTool search volume data – but this looks increasingly unlikely, which is a terrible shame.

Great post David

Cheers Carla – it appears this is a hot topic at the moment!

yes, you’ve hit a hot button – at least for me.

as to the tool being changed significantly in the final 2 weeks is, i imagine potential for that is near 0%. imo.

and unlike the changeover to new AdSense interface where they had a ‘feedback’ button, there has not been one with Keyword Tool.

i finally resorted to just posting in the help forums cause i had no idea where else to do it.

http://www.google.com/support/forum/p/AdWords/thread?tid=2b1ad4bf69fdcf86&hl=en

as far as i can tell, there has been little/no dialog on these forums about the deficiencies of the new tool.

Great article David, thanks. Don’t know how involved you get with PPC, but adwords can be equally frustrating.

Hi,
For Top Keyword, I check with Google Insight:
http://www.google.com/insights/search/#q=Fish%20Pond%20Supplies&geo=GB&cmpt=q
And I compare with another keyword already well positioned and that I track in Google Analytics.

Thanks for that I have always been suspicious of the results from the adwords keyword tool, it always seems to have batches of words all with the exact same amount of searches, it also doesn’t seem to differentiate between word order.

David what kind of CTR did you see for the keyword in Google Webmaster Tools? Did you see similar click numbers there or were they also far off?

Its awful if its that far out as people spend a lot of time and money designing sites based on those keywords.

Great post.

I had this happen on one of my websites the other day and it’s frustrating to say the least and I was too busy cross-checking previous keyword analysis to post…

Was there a consensus on which keyword tool was the most effective for seo keyword research?

I’ve been cross-checking the Adwords tool with Wordtracker and the SKTool. Personally right now I feel Wordtracker is the closest to being accurate for organic search numbers.

This sounds really familiar for some reason.

http://smackdown.blogsblogsblogs.com/2008/07/15/why-the-google-keyword-tool-is-useless-for-seo-even-with-exact-numbers/

Oh Crap….

I just went on a web sight buying/hosting binge after doing research on Adwords… On closer inspection with a rival product the results dont look 1/2 as good…

I suppose there is no such thing as a free meal after all. Where were you last week dave!

Keep up the good work mate

I stuck my toe in the affiliate marketing world 2 months ago. Before that all i could do was get email. I’m glad I came across this post when I did. I feel duped. I see that today more than a hundred people are protesting outside Google. “Do no evil” indeed. Thanks for this.

Spot on !

I used to rely on wordtracker, then on the adwords tool and now?

Not much to rely upon, what seems to work for me (and my clients) is using a combination of tools (traffic estimator included) and some common sense…not an exact science for sure and…very disappointingly a bit misleading

1 Phrase? Irrelevant data. Basing visitor activity beliefs around suggestion tools? Waste of time. Not providing specifics about Page Titles and Meta Descriptions used and as compared to competitor Page Titles and Meta Descriptions? Flawed. Thinking worldwide data suggestions should translate into local specific search? sad.

I’m glad this is getting a lot of coverage today. Both tools sucks. The SK Tool is showing me a term has 230 searches per month. I rank #3 and get 50 visits a day from that keyword. That’s a pretty sweet CTR.

My answer was too long David. I wrote a post instead

Why (And How) The Google AdWords Keyword Tool “Should” Be Used For SEO

http://www.seobloom.com/article/574/why-and-how-the-google-adwords-keyword-tool-should-be-used-for-seo-a-case-study/

Why do you not run an adwords campaign with the exact keywords you are interested in, you should see how many impressions the phrase gets.

@David In theory, the Google AdSense Search network should only cover the search sections of those websites. If the Google AdSense Search network is the only explanation for this major difference in keyword traffic (you said these sites may be manipulating AdSense by pretending ads are served for a keyword search), wouldn’t Google want to fix that?

If you are right, not only is the Adwords keyword tool off all the times, many Adwords advertisers are paying too much for showing their ads. And we wouldn’t be talking about a small margin, looking at your data.

Have you contacted Google about this? Considering the nice discussion you started with your article, maybe Matt Cutts has something to say about this?

@Jaan – I checked that yesterday, but unfortunately it seems it was before Webmaster Tools displayed impression/click data.

@Sint – no I think you are right, it probably doesn’t explain the whole inaccuracy. I’m not sure what would.

@Alan – sorry don’t really follow you there…

@Sandro – agreed, common sense is often the way forward – especially with long tail traffic.

@Bryan – interesting point, worth doing a blog post on and sharing it with us perhaps? I would be interested to see an example that shows the SK Tool being inaccurate.

Hi David. Thanks for sharing the post – sounds like that kind of *system bug* for data forecasting with Google tools could cause some wayward predictions (or worse still – business plans)

My main take-out from your blog post – be careful when building out forecasting data from a single data source. A quick resolve would be to cross check the Google Adwords data with another data set (I love Google Insights for Search) and you should then be able to quickly validate where Google Adwords has thrown you a curve-ball (any short-term discrepancies in search volume from Adwords should be obvious alongside Insights for Search).

However, Google Insights for Search won’t give you monthly exact-match data so you’d need to extend a large chunk of assumption to compare the trends of the two data-sets OR cross-check your data on phrase and broad match and then do your own extrapolations for exact-match.

And if by chance – I wrote a blog post over the weekend looking at the accuracy of data from the Google Adwords keyword tool versus Google Insights for Search. The two data sources both offer good value to keyword benchmarking and forecasting, although Google Search for Insights data (with a little work) offers better reward. The blog post has just gone up – http://www.coastdigital.co.uk/blog/2010/08/17/google-adwords-vs-insights/

this article’s bang on, we have multiple client sites ranking #1’s for big traffic terms, and the actual hits are about 1/20 – 1/30 of what the adwords tool implies you should see (calculated using 40% for a #1 position) on the “exact” readings.

so taking matt’s example “Tool results of 15,000+, real world of 200ish”

15,000 x 40% = 6000 ? actual = 200. 6000/200 = 30x optimistic.

only I thought everybody knew this already?

David,

I tried to replicate your search in the keyword tool and actual found it to be fairly spot-on to the SEO results that you found. Assuming you chose your location as United Kingdom and language as English, the local monthly searches for [fish pond supplies] came to only 210. How did you get +18,000?

@David – Ever seen a 866% CTR? http://www.bigleapweb.com/blog/best-place-for-google-search-volume/

@Robin – I don’t trust Insights for Search much either. It is showing a keyword I rank #1 for is a “Breakout” search. I maybe get 1 visitors a day from that term, at most.

@David (the David 2 comments up), are you looking at [koi pond supplies], the phrase right under [fish pond supplies]? I see that one at 210 searches.

@Aubrey – I just double checked, the GA Keyword Tool shows around 18,000 and it is definitely set to exact match and the UK.

@Bryan – nice post

Nice share..

I think you have a point here David. Google Adwords tool has always been on my suspicious list because it shows exactly different statistics when compared with http://www.wordstream.com/keywords/

But I still use it to get an overview about a certain niche and how deep the market is. Then after wards, I employ other keyword research tool to help me out. But thanks all the same, I know better now.

Hi David,

great blog post. Goodonya! What a coincidence, me and a friend of mine have just experienced a similar difference between the Google Adwords Keyword Tool and the SBK Tool on a high volume top keyword 2 days ago. Eversince I’m trying to find an explanation for the inacceptable difference between the 2 tools.

The Adwords Keyword Tool is telling me that the exact match has a global search volume of 5 million/month whereas the SBK Tool is talking about 3,500 searches/month only for the exact same keyword. WTF????? Fortunately I have the keyword running on exact match in Adwords and the average impressions for this keyword for the last 3 months has been round about 250k. (I’m in the German market by the way, where Google has something like > 90% market share)

Talking about Adwords there are some factors that have to be considered like impression share (depending on budget and position) and so forth, so maybe the actual search volume is a bit higher. However, my example clearly supports the argument that neither the Keyword Tool nor the SBK Tool show accurate figures for us marketers.

So if any one finds an explanation I’d be thankful to hear it

Meanwhile trying to discuss this with a couple of my contacts at Google. Anyone else trying to get an official statement?

one more add to my last comment:

even the local search volume from the Keyword Tool is far away from the truth with 3,5 million searches/month

@Matze – thats interesting. Those impressions is it just Google Search Only or is it the Google Search network?

The search volumes that are being projected are based on a previous month’s data. While being 99% off is no slight digression – have you factored in trending keywords searches?

I’m noticed that the SK tool doesn’t show data for some keywords. For example, I played around with “termites” and “termite control”. The Adwords tool shows 74k monthly while the SK tool shows only 35k. Great, seems to follow the point of the article.

The issue was when i searched “termite control”. No results in the SK tool. How should you estimate realistic volumes for keywords that don’t appear in the SK tool?

Possibly there is a multiplier to where the SK tool usually shows X % of the AdWords tool. Anyone run any analysis on that idea yet?

good point – the Adwords numbers refer to the Google search network

@David – the volume has been pretty much the same for the past few months, in fact April was the peak month according to Google Insights I believe. It might not be like this for all keywords – but that is what makes it so untrustworthy, especially if you are basing a business plan on it.

the best use of this data is to show the order of importance of keywords and variants rather than an absolute demand.

As a domainer I use the Google Keyword Tool every day, but always with a grain (or a shovel load) of salt. I’ve developed many keyword domain sites that the keyword tool represented with “thousands” of “exact match” searches. My traffic % from Google organic results for these sites (when on page 1 of Google search results for the keyword phrase) mostly backs up your data.

When I look at Google’s exact match results I immediately take into consideration the potential traffic I could also receive from Yahoo and Bing organic results (usually a combined 8% – 12% of all organic search traffic for my sites) plus all of the other related search phrases that will result in Google organic traffic that I am not necessarily targeted through my content or the exact match keyword domain. Point in fact is that one keyword domain site now receives organic Google traffic for 375 keyword searches…all related in some form or fashion to the keyword domain.

It’s a trade-off, but as long as you know how to decipher or account for Google’s “inaccurate” data you can still use it in your decision making process with some level of comfort.

Same as George stated, some words have no results in the SK tool, like “web host reviews.” Please tell me this doesn’t mean zero.

Great article,

All of us are pretty much held hostage by Google. The only ones who know true traffic values are Google and the great black box. As much as SEO’s are loath to use ADWORDS. Its the only way you’re going to discover what works for your particular site, geo-target and market. Don’t you think its better to make sure your quality score and marketing are up to snuff, before you plunk down big money on an SEO initiative.

Rand Fishkin “The Art of SEO” says that PPC traffic accounts for about 8% of traffic click thru’s, Positions 1-3 are equivalent performance to positions 7-9. (Disclaimer – I’m an Adwords guy)

So do the math, factoring in impression share, your best 20% of keywords are your winners multiplied by 9 should be your traffic at position 7-9 front page. Pretty straight forward.

Throw in Google Analytics to verify your results, now throw that stuff into your favorite tools for traffic estimates. Feel free to share your methods.

Searchengineman

David, both tools (search based and the old keyword tool) are giving wrong estimates. When doing keyword research for SEO purposes, we choose the best ones by comparing them and we rely very less on search volume figures. If we have access to AdWords history we run a keyword report (google searches only + exact match) to get a closer to reality idea. Then we take the into consideration the publicly known CTRs for each position in the search result pages and we come with our own estimates. To summary, the closer to reality results are coming from AdWords itself and not the free tools. You pay for the keyword and you get better details.

Also, choosing a keyword just based on search volume is simply dumb. Create a list of what you think are the top 10 choices to run after with SEO, then test them (AdWords), look at conversion rate, bounce rate, landing pages (use Google Analytics) and only when you find the perfect combination of CR+traffic values you have a winner.

I wonder, for the keywords that Google SK tool is not showing whether it’s possible to compensate that with Google Insights, so if google is showing 3k searches for ‘seo tools’ and you want to see the traffic for ‘search engine optimization’ tools then is it possible to go to insights and compare those 2…

@Darko – good idea, @emory check out what @Darko suggested.

ITs been interesteing reading this. I too get the same thoughts when the exact searches are so high and even after rankin on top 5, I hardly get 5% traffic for the site. But stil it would be useful is someone did mention a more precise and accurate tool. Cheers

nice post.
i’ve had similar experience already.so i said to everybody around me that this is only a reference.

hey,none of you guys seemed to remember search trend,using Google Insight to find the search volume high and low through seasons.

when you want to find how much traffic a phrase is going to bring to you,G Insight is a must.

@Daniel – there isn’t one. Think about it, who is going to provide more accurate data on Google than Google?

Case study.

Keyword phrase “starcraft 2 hotkeys” has a global monthly search volume of 1600 (exact match), and they don’t have any data for local US search volume. (According to Google adwords keyword tool)

So I guess I shouldn’t receive many clicks for this keyword right? I’ve received 10,287 visits from this keyword in the last 30 days and 5,980 of these visits were from US.

This is just my personal blog, while working for many other SEO/PPC clients, I find Google adwords keyword tool is incredibly inaccurate.

@Rui – interesting, I used to run a Modern Warfare site and I got around 10k uniques per month for “modern warfare 2 tips” – but again, the volume wasn’t available in either tools. It was however a breakout term on Google Insights, so perhaps it could be to do with how long the keyword has been about? You can’t predict something on less than 12 months data really – because of seasonal variations and spikes such as “BP oil spill”.

I know the keyword tool reports “searches” not “clickthroughs”. Plus I had read somewhere from Google that only half that searches result in clickthroughs. So I have been automatically deducting that which seems about right for my normally obscure technical B2B clients. Is it possible that some search term numbers are influenced by all the optimizers checking their rankings? (in addition to all the normal error that occurs for all online measurement) It seems like the numbers get more weird for some competitive terms, and Google’s reporting gets more weird for some categories of terms that are more hype-related. I know that the local numbers tend to be off even more than the global – sometimes reporting very low numbers for USA English where there are thousands for the global, but if you look at all the main countries that speak English their numbers are all proportionately low as well. There are also sometimes disparities in traffic month to month that makes no sense whatsoever.
Looking at some main travel industry search terms for one client I am getting about 1/5th the traffic for a term that I would expect, given a certain ranking. Could seos be multiplying the number of searches over clickthroughs by 5x more? Seems possible to me.

@jebbiii – The CTR for a website in the top spot is approximately 40%.

Interesting article, I have many similar results. But also, on the flip side – keywords with ‘Not enough data’ have proved to be anything but low search volume terms!

I have been using the Adwords Keyword tool for all of my keyword research. Unfortunately it is not good for predicting organic traffic but it goes give a basis on which to create a strategic keyword plan. I only use exact match and make sure to specify a location. The numbers don’t mean much really unless you plan to spend all your money on Adwords.

Remember Adwords is how Google makes money – in fact it is almost their sole source of revenue. So I don’t think they have an incentive to distort the data. I haven’t found other keyword tools to be effective in giving me an idea of what keywords to focus on. I don’t use the numbers for traffic estimates but only to gauge interest.

I hardly ever use sktools but looking at the help site, they *sorta define the difference between the sktools and the adwords keyword tool:

http://www.google.com/support/sktool/bin/answer.py?hl=en&answer=114875

Both use different calculations to come up with their numbers. I don’t know who/what to trust. So I use several sources for doing keyword research.

this article is really useful to me, i am starting to build up my own website, and thought a good idea would be to use the keyword tool in order to use the most popular to build new blogs that would give 1 way links across to my main site.

i am now having second thoughts of the usefulness of this tool.
many thanks for the advice.

You have a valid point in that you shouldn’t be using the Google Adwords Keywords Tool for SEO purposes.

But the Google Adwords Keywords Tool was never meant to be used for SEO; it was for finding keywords for Adwords! And the numbers shown are the maximum number of impressions you would get a month (across all properties) if you ran your ad for that keyword, and bid something stupid like $99.

Still, I can’t believe that only 66 people search for “fish pond supplies” every month. Something is screwy.

Kinda stating the obvious with this linkbait – if people do not understand how to use it or how the results are calculated, and many people are hobby seo’rs with barely a clue, then this is their fault.

It is actually an excellent tool for keyword research – just ignore the numbers.

Oh thank you for your great insight for this article about google adwords. actually i am using it right now and now i am kinda checking the Search-based Keyword Tool of google that you point out at your post.

Thank you for this David. It might help me in my keyword research today..

I agree that the Google key word tool seems to work better.

I use adwords keyword tool a lot but always take it with a pinch of salt, and use other tools aswell. This is one of the most difficult things to prognosticate for a client – always get asked what traffic will I get in organic compared to top 3 adwords position. I generally use adwords keyword bidding to find out if the keywords ar worth optimising for in organic.

Oh thank you for your great insight for this article about google adwords. actually i am using it right now and now i am kinda checking the Search-based Keyword Tool of google that you point out at your post.

Thank you for this David. It might help me in my keyword research today..

The main problem here is not so much the tool but more the expectations around the traffic you’ll generate based on the tools reported search volume. Regardless of the traffic you receive relative to what the tool reports, it’s still useful for gaining an understanding of how important one keyword is vs. another. At the end of the day if you have a huge site which is pulling in traffic for millions of terms, gaining an understanding of their importance based on search volume is invaluable for prioritizing how you optimize the IA. We’ve been using the Adwords API to pull in search volume for years to sculpt our IA based on search volume and it’s worked incredibly well for us at Expedia…

David,

I just checked the keyword I’m trying to rank for through both the AdWords Keyword Tool and the Search-based keyword tool.

The only weird thing is with the AdWords tool (exact) it says 12,100 searches local. So I checked the Search-based tool and it said 18,000 local.

I don’t get this? Looking at your results, fish pond supplies got 18,100 from the AdWords tool and only 66 from the Search-based tool (which I just checked and it says 440 not 66) which is 1/274th of what the AdWords tool says but for my keyword its the Search-based tool even gave a higher number than the AdWords tool.

Does this mean if I were to rank for this keyword I would see some pretty decent traffic or do you think it’s wrong?

The reason I ask is because I’ve been working on ranking for this keyword for 3 months now and I don’t want to get to the 1st page and it all not be worth it.

Thanks,
Caden Grant

On 8/27/2010
I just check for Google Adwords again and it’s change

[fish pond supplies] 1000 880
fish pond supplies 22200 18100

So google see this article and check it’s bug?

@Andrew – very good point.

@Caden Grant – That sounds interesting, what keyword is that?

You have to login to choose a country – I’m based in the UK, if you check the UK search volume for the SK Tool it is still 66. It is 440 if you don’t login – I think that might be global.

It sounds like both the AdWords Keyword Tool and the SK Tool are saying it is going to be a good earner, so I think you’ll be ok.

@cc-idea I’m still getting it at 18100 for exact match – perhaps the Google AdWords tool I am using has a bug? I think there are a few versions – but they will be combining them shortly.

Which also goes a long way in describing why you should never write blogs or articles based upon what might be searched for according to any source. But you should work based on your passions because THAT will be far more fun than any search result for a few months.

Trends come and go. Christmas blogs are most relevant from October and until December 24th, whereas swimsuits don’t sell all that well in cold Scandinavia whereas it may be relevant when March comes when people begin planning their summer vacations.

So you can search optimize your passions, but can never build a passion upon search result numbers.

Worth remembering, I think.

I think so, i also face the problem like you. I am very disappointed.

I find the results to be out dated. If you are an owner of a website that is supposed to be up to date and current, you really do not want to use the adwords keyword. I run a grand theft auto website and star craft website. As the release of the latest starcraft happened I did a check to see what people were looking for , I found only 14 people a year searching for starcraft. When starcraft ii sold in excess of $1b in the first day.
the 14 search statistics can not be correct.

you are all wrong, even the author. Don’t any of you pay attention to the fact that the beta version of the tool, offered a few months ago, that is now retroactive, is what is skewing your numbers? If you want any type of real accurate numbers, you will have to use the previous interface that is in the upper right. Not only that, but the keyword phrase fish pond supplies is very general. You cannot compete with this search term easily because it is not a narrowed down niche keyword. Yes it is not a single term keyword, but you are still “fishing” around with this keyword phrase. You need to drill down anyway and find a more targeted niche keyword phrase that will most certainly convert better as well. Who cares if you can rank someone for fish pond supplies when the conversion rate is paltry. I would rather have a lower visitor per month keyword phrase that converts to more sales any day of the week.

Your information is misleading and inaccurate and your example does not provide a sense of confidence in your skills as a supposed expert on SEO.

Maybe you guys should be more aware of newest changes before making statements that are flaccid and incorrect.
check twice then write a post about it.

Hey Carl,

The AdWords keyword tool has always been grossly inaccurate. Even using the old version it still says 18,100 searches – which is the same as the new version. Plus I did the research back in March and I believe it showed something higher back then.

“Fish pond supplies” isn’t general – if someone wants to buy multiple items for a pond, then it is actually quite targeted. The site in question was limited to the software so I couldn’t track ecommerce properly, but I could track the number of people who added to basket and got to the page before checkout. Looking at June, July and August they have had 50 visits for the keyword “fish pond supplies” of which 6% have reached the checkout page before payment. That’s not exactly a low converting keyword is it?

The information I have placed in my post is not misleading or inaccurate, there are screenshots clearly displaying the numbers.

I’m aware that the keyword tools are all going to be combined shortly, I am keeping my fingers crossed.

Oh and Carl, you might want to sort out your duplicate homepage issue:

http://howtomakercairplanes.com/
http://howtomakercairplanes.com/htmrca-home.html
http://www.howtomakercairplanes.com/
http://www.howtomakercairplanes.com/htmrca-home.html

Perhaps use a 301 redirect and a REL canonical just in case?

And I would remove that sitewide nofollow to your homepage on your blog – kinda stupid when its linking to your own site.

oh my word one would think Carl wrote the tool or something?

..to come onto a high level seo blog and show that much ignorance at the tail end of such a comprehensive post and comment trail..

and when you clearly have so little clue about, well anything really..

foot – bullet – ouch – doh!!

Great insight. Any suggestions about using the Google Search Based Keyword Tool when the site is not associated with an AdWords account? Thanks!

Nice, you wrote about Google Adwords keyword tool. But, when you are targeting a keyword should have some weight and as the same for your Domain too. You told that that keyword which is in the 3rd place(good position), in which search Google.com, Google.ca? etc…! Your keyword should have more than 80,000 to 10,0000 search volume of global and 30,000 of Local search volume. It shows in position 3, but in other Countries it will show in the 5th page, or not visible. Some SEO and link building is required too.

Hmmm… the basic idea I received from the very best SEO people here in Denmark was: start with an Adwords campaign and you will gather so much valuable information regarding keywords – now with your added perspective David – and the lack of trafic which keeps starring me right in the face after keywordoptimizing using the Adwords keywords tool – I definately think there is a lot of SEO guys out there that need to read your article. I think I will write on it in Danish on my private blog and graciously extend a link to you.

Thanks
Anders

Cheers Anders

Great post David – I’ve been using the AdWords Keyword Tool for about two years and always relied on it as a “relative” indicator of keyword search volume. I still think it’s the best thing we’ve got, and I’m incredibly attached to the forcelegacy=true version (in fact that’s my home page in Chrome).

I was really saddened when they updated the tool and all the local volume numbers went caput, with just about everything returning “Not enough data”. Thank God I saved my old numbers and can still return to those… I just don’t trust this thing anymore.

I ran a test with some of my own terms to match your own and saw the exact same thing. Tested multiple keyword silos across three different verticals (hotels, pharmaceuticals, and healthcare) and saw the same distortions between the AdWords and SK tools.

The simple reason Goolge inflates it’s numbers: They want more money from AdWords Advertisers. They always have, and always will, be chasing the eternal $$$.

Do No Evil? More like See No Evil.

Wanted to jump in here because this post is leaving a lot of FUD, and causing problems for us

We’ve spent a lot of time and effort analyzing the accuracy of keyword data (because we have 197,000 users relying on us to give them accurate data via Market Samurai).

There is an issue around Search Network data being included – but it’s not as bad as is suggested in this post.

Instead, I think what you’re seeing here is a sampling issue.

If I review the data for “Fish Pond Supplies” (exact match, UK, English), I’m getting this data:

– Search-Based Keyword Tool: 66
– Beta Google Keyword Tool: 210
– Legacy Google Keyword Tool: 14,800 (Note: July data)

Obviously not everyone who searches for a keyword will go to 1 site.

So if you use the AOL CTR data to work out what traffic you should get from this…

…Assume 54.5% of people click ANY result…
…Assume 8.5% of them will click #3…

8.5% * 54.5% * 210 = 10 clicks.

10 clicks is very close to 12 clicks.

Based on this, the Google Keyword tool comes up trumps. (Compared to an estimate of 3 clicks from the SKTool.)

(Run this on a few more keywords and you’ll see that the Google Keyword Tool is actually quite accurate. This is why we use it inside Market Samurai.)

We often get support tickets where people have seen one set of figures one month, and a very different set of figures the next month… the “gold nugget” keyword they’ve found turns into something worthless, and they feel cheated.

So we’ve had to analyze literally hundreds of keywords, like this.

I believe what people are seeing here are sampling issues.

Obviously it wouldn’t make commercial sense for Google to count every search through a turnstyle every month. With 88,000,000,000 searches a month, you’re looking at a HUGE amount of computing power.

What they are more likely doing is taking a sample (a few percent) of searches, analyzing this, and coming up with “Estimated” searches.

But even with 99.9% statistical certainty, 1 in every 1,000 keywords will show “bad” data.

(The percentage of “bad” keywords is probably higher – maybe as high as 1%, although I’ve never gathered a few hundred thousand keywords and watched them over the next few months to know for certain.)

This is why I believe it’s a sampling issue, rather than the data coming from the search network.

If the problem was a sampling issue, we’d see a spike one month, and then the next month (when the data was re-sampled) the figure would be back within reasonable bounds (“fish pond supplies” is now back to 210 per month – and I can point to a dozen other examples from our users.)

If the problem was just the search network, we’d see inflated searches on product-specific keywords (Amazon is part of the search network) – but we wouldn’t see temporary spikes like this – the traffic would be more consistent.

If the problem was sites like Mahalo redirecting traffic to different search queries, we’d expect that these redirects would probably be applied manually (otherwise, why not just use Google’s own ad matching algorithms). And if the redirects were applied manually, we’d expect that the redirects would be permanent (i.e. no spikes), and would affect only a small percentage of high traffic keywords (i.e. would be big, and rare, to maximise ROI).
Big and rare variances I can agree with – but in this case, the next month, the figures were back to 210 searches. Permanent? No.

On balance, I don’t see how this can be caused by the search network.

There’s every chance that the issues above MAY be affecting the data. And you’re right – there are some issues with using the Google Keyword Tool data for keyword research.

But I don’t think they’re the cause of your specific issue.

In this case, a sampling issue is just the simplest explanation.

I stand by the Google Keyword Tool – it’s still the most accurate source of data around (as I demonstrated with the 10 clicks estimated vs 12 clicks actual traffic – pretty close, I say.)

I’d love it if they gave us 100% accuracy on keyword data – but I understand it’s not feasible. 98, 99, 99.5% confidence on the largest sampling made available and the most accurate data available – I can live with this

Brent

P.S. – Feel free to email me if you want more examples / more specific data.

I’d like to thank you for this article as well as warn you about a limitation in search based keyword tool. It looks like sktool is censoring certain words but the principles are unknown to me.

For example, it doesn’t accept any phrase with the word “nintendo” in it. This word never appears in the results either but it has no problems with “sony”. For someone with a gaming blog like me, it’s very frustrating to not even know the logic behind it. I can’t really trust ANY keyword tool now can I?

Great article. This is a VERY serious issue. What I see is this.

Beta keyword tool – very low “exact” search volumes

“previous keyword tool” – show high “exact” search volumes in comparison to the exact same keyword phrase when using beta. Example, one keyword phrase showed me 550,000 exact search on the old tool, and the new tool show around 1000. That’s crazy.

Right now, no word from Google. In my opinion, the keyword tool is BROKEN. Data I have from the past year said one thing, and the beta keyword tool completely contradicts my previous findings.

My advice? Email or post on the adwords google forums. Ask about these crazy differences. I’m not overstating the impact on people out there using the tool. Beta vs. old shows completely different volumes. People are basing financial decisions on what appears to me flawed data.

Has anyone heard anything official from Google? This is a complete nightmare for someone like me. It ruins investments I’ve made and planned to make. I don’t know up from down and left from right. The data is completely unreliable. In reality, the tool should be disabled until these discrepencies can be addressed officially. I used this tool religiously over the past years. It’s just an awful feeling right now. I’m sure I’m not alone in this dark, knotted stomach kind of feeling about data you have been basing your decisions on.

There’s a thread about this at with regards to huge discrepancy between old tool and beta version

http://www.google.com/support/forum/p/AdWords/thread?tid=5241a33035617b73&hl=en&fid=5241a33035617b7300048f9b5f546cab

I also feel sick to my stomach, not just thinking of myself, but of so many people in this industry who have relied on this tool. Its pretty disgusting. The differences in the results between the new beta and the old adwords tool are just ridiculous, especially on exact match. I have spent several hours researching this and I am still pretty lost. The only conclusion I’ve found is that you HAVE to spend money and run the words in Adwords to get the impressions, the real data. Big unhappy face =(

Hi Brent.

interesting comment, so you’re basically saying to multiply all readings by a factor of 0.54 to start with, and then again by the factor relevant to the ranking position and the new adwords tool results will be close?

still not very (close) on the several client rankings I just checked, we’re still getting approx 1/4 to 1/3 of what it should be, and that’s allowing 40% clickthrough for a position #1 & 2 indented listing.

it is however much better than the 20-30x factor from the legacy tool, so its a start.

@Kev – 0.545 is essentially the percentage of people who do click. (Not all searches result in a click.)

Make sure you’re looking at the right data (by location). i.e. If you’re ranking in the USA, look at the local traffic for the keyword for the USA. This might be the cause.

Hey guys, I’ve written an updated post, you can find it here: http://www.davidnaylor.co.uk/is-google-adwords-keyword-tool-now-more-accurate.html

@ Brent

nope, definitely UK, exact, English. even using x 0.54 first, the figures say 3 -4 x the actual hits as per analytics. happy to share exact examples offline if you’re interested.

& as I say, this is not one isolated case, this is across the board on large client sites with multiple (allegedly) high traffic #1’s – #3’s.

they do bring lots of visits, just nothing like the tools say they should.

This is exactly why PPC is so important to test out these theories before the benchmark keywords are commited to. Just recently a client had his SEO chasing a benchmark for 6 months that supposidly drove 14000 visits and now pos#1 reality bites and receives 430 avg….6 months ouch thats gotta hurt!

Great post BTW

The problem is the google adwords network is made up of loads of different search engines, and google collects the data from those engines. The google data makes up a small percentage. Its best to use the google search base keyword tool because its deals purely with google search data.

A google employee has responded to this issue at

http://www.google.com/support/forum/p/AdWords/thread?tid=5241a33035617b73&hl=en

Apparentely as mentioned previously in this thread, he says the old tool takes it’s data from google and it’s search partners where the new tool is just google

A link was posted here, but this is the acutal comment from Google. This is incredibly important so read it!

———

A colleague spotted this thread a short time ago, and asked if I would post this information by way of clarification:

“If you use both the previous and updated versions of the Keyword Tool to search for keywords, you may notice differences between the tools for statistics on Global Monthly Searches and Local Monthly Searches. This is because the previous version of the Keyword Tool provides search statistics based on Google.com search traffic and traffic from search partners, while the updated version of the Keyword Tool provides search statistics based on Google.com traffic only. We’ve updated these statistics based on user feedback, and hope you find them helpful for keyword selection.”

Regards,

AWP

——–

Every time I click on the link it opens me a blank page. What could be wrong?

What link would that be?

http://www.google.com/sktool/

It shows me a blank page.

When I look at it, it shows a notice that it is going to be discontinued soon. So I guess it has been discontinued in your area.

I’m completely lost with all this now! Not sure what data you can trust. I’ve been using Market Samurai for over 2 years now and they have just updated to use the new Google Adwords search tool and it offers a keyword search without logging in to your Adwords account (100 results max) or (800 results max) if you log in. Unsure of what figures to use now?

Hi David

So glad i stumbled across your article. Very interesting. Sounds like i’ve got more chance of getting a snog off Cheryl Cole (Tweedy) than getting even near the 1st page of Google!
I’m article writing, youtube videoing, commenting on running and jogging articles and i’m not comming up in search much………..Somebody pleeeease help me!

Thanks for your article tho!

Thanks for the post.

After an unsuccessful campaign I’m trying to get as much useful info as I can to improve the second one.

I see the search based keyword tool will be discontinued. Do you still not recommend using the google adword tool? It says it is/will be improved.

Tommy, as it says at the start of the post:

“Update: The AdWords Keyword Tool may now be more accurate, please see our latest post.”

The new version is much more accurate.

Interesting. I haven’t ever bothered with the keywords because i didnt think they did much for me anyway.

Great post! I think we all had huge doubts. I’d rather do my SEO based on a (short) AdWords campaign.

Great feedback and explaination of technical results, I will try some of the tools mentioned to help me work through my keywords and text – previously I’ve used the Google Adword tool to build these for my website http://www.ECO-Pressure-Clean.co.uk which I am trying to get placed in a 1st position in Google for Driveway Cleaning Edinburgh – now in 2nd but want more !!

Keep up the good work David.

i am a complete novice at seo and website in general. Times are hard i needed a website could’nt afford one could’nt afford to be without one. So learnt how to how to use dreamweaver via youtube videos lol. I need to also know about sep etc so searched and gathered a lot of information. Now i have read this article and it started saying google adword tool was no good by the end of the article it said it was ok and to give it a try. So it would seam no matter what you do it will change to suit eventually.

Here is something that should cheer you up a bit

Last year I was doing some research using the Google Keyword Tool and found that searches for motorbikes in my area were being searched for 200k+ per month.

With that I decided to go into the motorbike business and fill the massive demand in the area. The website was completed and within a few weeks was right at the Number 1 spot in google for the 200k+ pm search term. I really thought I had cracked it and waited for the orders to come rolling in. But they never did!

Instead we got about 16 visitors a month at the number 1 spot. I was gutted and looked for answers. I found the SKTool and it showed a far more accurate figure of 22 visitors per month, hehe…

Lesson learnt! The Google keyword tool only shows how many ippressions a google adwords ad could expect to get for a search term, it has nothing to do with SEO!

Sadly Google have now discontinued the SKTool so I guess we are just all in the dark. I say just choose search terms that make sense and get your site to number 1 and just see what happens… :/

This is a timely post. I’ve literally just started using keyword tools. Much to my relief it appears that I accidentaly have been using the Search Based Tool.
I wsn’t sure I was doing this right, but your post has made me more confident.
Many thanks

Alan

The Search based tool seems to have been replaced. You get redirected to the Adwords tool and I cannot see how you can get results without the content element now. Any one have any suggestions?

Hi David,

Great blog love your slogan about a good rank!

when do you think we will see a better and more useful tool available to all us seo geeks?

thanks

Useful article. I don’t think one should rely on any one tool for keyword research. If you are planning to invest time and money on niche keywords, ensure that you double check search volumes via at least 2 different tools.

I agree with you .Its not helpful for our site.

Why The Google AdWords Keyword Tool Shouldn’t Be Used For SEO – A Case Study

158 Comments

Get in Touch